Antiretroviral therapy as a risk factor for chronic kidney disease: Results from traditional regression modeling and causal approach in a large observational study

OBJECTIVE: We investigated whether patients receiving selected antiretroviral combinations had a higher risk of chronic kidney disease (CKD) using traditional regression modeling and a causal approach in a large prospective cohort. PATIENTS AND METHODS: For the purpose of this study, we selected 6301 patients who (i) started their first antiretroviral regimen after 1(st) January 2004, (ii) had at least one serum creatinine measurement within 6 months before ART initiation (study entry), and (iii) had at least two measurements after study entry. Baseline eGFR was defined from the last serum creatinine measurement before study entry. All eGFR values were calculated using the Modification of Diet and Renal Disease (MDRD) equation. Both traditional Cox proportional hazards model and Cox marginal structural models were applied. Distinct coding for antiretroviral therapy exposure were investigated as well as double robust estimators. RESULTS: Overall we showed that patients receiving tenofovir (TDF) with a ritonavir boosted protease inhibitor (rbPI) exhibited a higher risk of CKD compared with patients who received TDF with a non-nucleosidic reverse transcriptase inhibitor (NNRTI). Such an increased risk was observed considering both initial and current regimens. Our analysis revealed a clinician-driven switch away from TDF among persons experiencing a decline in renal function while receiving this drug. CONCLUSION: Our results show that combination of TDF and boosted protease inhibitor is associated with a higher CKD risk than TDF and a NNRTI.